Did a test after noticing that Hot Streak proc chance was on the low side. The test was done on a level 85 mage, 24.26% character crit rate, talents where mangled to not include Impoved HS.

Casts: 1435 (scorch)
Hits: 980
Crits: 455 (31.7% fits the 24.26 + 5% from critical mass)
HS Procs: 119 (26.15%)

The previously known linear approximation is -1.7106 * hot_streak_crit + 0.7893 which expects 37.43% hot streaks or 170 procs. That is a very significant change or namely at 24.26% crit chance the proc chance was reduced by 10% from the previous beta values.

Regretfully I don't have accesss to a level 80 mage to check all the values, however if anyone else has, that would be appreciated.

One other thing, I have done some tests on the Pyroblast glyph, and it looks like the glyph doesn't really have any effect on Pyroblast! at least not on the dot ticks generated by it. I'm still checking the numbers in the logs.

Looking at your world of logs you linked tyrian, it appears pyroblast with it's dot is more damage per cast time than fireball and it's more dpm than fireball even if you spam it. It appears it would be worth casting pyroblast if pyroblast! dot isn't ticking.

Have you guys tried modeling/simulating pyro as a nuke instead of just an opener or proc?

I was a bit surprised by the result but you are actually right Mentalfloss, hard casting Pyroblast when there is no pyro dot on the target results in a DPS gain (~1% for BiS ilvl 372 profile).

I added the following line to the priority list:

pyroblast,if=(target.time_to_die<60|mana_pct>39)&!(dot.pyroblast_hs.ticking|dot.pyrob last.ticking)

Which basicaly means : if the dot from Pyroblast or Pyroblast! is not ticking, replace the fireball you were about to cast with a hard cast Pyroblast.

Here is the html output for detailed results: Simulationcraft Results

Edit: Take these results with a grain of salt for the moment. There might be a bug in Sim Craft that creates this DPS gain. See Cataclysm Mage Simulators and Formulators for details.

Interesting indeed. However, even though a hard cast Pyroblast is more DPET with the full duration dot factored in, this tactic loses value if the DOT is overwritten by a subsequent Hot Streak Pyroblast!. The question is how long does the DOT from a hard cast Pyroblast need to tick to be worth the same or more DPET of a Fireball?

With the currently low Crit rates it is less risky to hold on to a Hot Streak until the Pyro DOT ticks down. Combustion Tracker will help keep tabs on the Pyro DOT.

I haven't left Exodar since Cata Launch. Currently I'm planning on doing more with Combustion and hopefully get to Ignite next monday, and then do some rounds on T3 HS. I was planning on doing a ~10%, ~20%, ~30% from gear test to show the line (4 hours of scorch spam per should do it), then a shorter test with Critical Mass to see if that impacts it.

I'll obviously be posting once I get the data together.

That would be extremely helpful Zaldinar, probably an up to 40% graph would be very good since that's the range most t11+ geared mages will be heading for.

I was looking over some things - and I've found something interesting. Correct me if I'm wrong please.

With the current value of "int" vs. "spell dmg" being at 4.2022 and 2.7381 I was looking over my trinkets.

Currently I have [Theralion's Mirror] which is clearly the best. It's a 359 item with a DPS proc and int. Hard to argue.

However, I have two others which I thought I knew which one was better, but now im not sure after doing through some things.

[Heart of Ignacious]

vs.

[Witching Hourglass]

Now the first one basically is giving me 385 Spell Power... which worth 2.7381 per point comes out to be 1054.1685 With this trinket I have 7041 SP and 24% to crit on my character page. The haste use comes out to be 267.5 rating.

The second is gives me 285 int... which worth 4.2022 per point comes out to be 1197.627. With this trinket I have 6949 SP and 24.46% chance to crit on my character page. The haste proc on this comes out to equal ~213.75 rating.

So using the relative values the epic trinket is worth 1459.29725, and the blue is worth 1521.351375 - making the blue quality one better.

HOWEVER

The epic gives me 7041 spell power - and the blue gives me 6949 spell power, giving that a difference of 92 spell power.

Int gives me crit rating, so let me include that in my thoughts -- the int gives me .46% - which comes out to be 82.4688 rating (179.28 = 1%). Okay so using this thread relative value for crit that 82.4688 = 204.522624.

So the extra 92 spell power from the epic trinket = 92. Using it's relative value it equals 251.9052. The difference between those two is 47.382576 in favor of the epic trinket. How if we count in the haste use/proc of the two as stated above the epic trinket has a better overall haste value by 53.75 - giving the epic trinket an overall total of 101.132576 over the blue trinket.

So pretty much what I'm saying.... is when you break down the trinkets the epic comes out to be better - using the math posted the blue is better. I believe int is rated to high. Unless having ~4000 extra mana comes out to be values higher than the number posted above - but I really do not believe that to be true considering with proper mana management that 4000 mana really doesn't matter - especially because we have a spell that doesn't cost mana.

Something doesn't fit - and I believe this math proves it. Comments - corrections would be most helpful - but I believe this is right.

Can you give a source for those numbers? Is this coming from simcraft, or vontre's or rawr or what? Incidentally, in your calculations you seem to have forgotten that the 359 trinket consumes the spellpower buff stacks.

I guess the reason is that more mana = more fireballs/less scorches = higher dps.

Yes we have a spell with no mana cost, but that spell also has lower dps than fireball which means we rather use fireball if we have the mana for it.

The source is this thread - and it came from simcraft according to it. And yes the stacks do fall off - but they're back up in a very short amount of time. That should be included you're right, but it still leans in the 359's favor -- even though with the straight up numbers the blue is ahead by a lot.

The numbers just don't match up, but they should. When I broke down the two trinkets if the numbers were right - I would yield the same things wouldn't I. It's like if I had 2+2=4. It equals 4, but if i broke that down to something identical and got it to be =5, something would have to be wrong. Maybe I'm not actually adding 2+2, I should be adding 2+3 and there was a miscalculation somewhere.

I'm just saying these numbers should yield the same results because they're comparing the same thing, but they don't. One view of it yields a different result of the other view, but they're the same. The relative numbers are the source for all of this, therefore it has to be that they're off slightly.

Basically I'm just personally trying to prove that it's correct for me to be using the 359 trinkets. My target dummy numbers are higher - and my raid numbers are higher - but "it" said it's not as good for me. Yes these numbers are slightly higher, but still higher. It also didn't "add up" in my mind when I had a trinket yielding 385 SP and a 20 sec haste use and not using it because I had a blue with 285 int and a lesser haste proc.

Out of curiosity, which races would be the best for each faction for a Fire Mage? I'm currently considering a race change for my Mage and I was curious as to what the optimal would be from a dps perspective. The question has been put forth by another poster in Simple Questions/Simple Answers; however, it is unanswered and seems to be from a more general point of view, whereas I'm looking for input on fire specifically. Thank you in advance for any input on the topic.